我有下面提到的不同的模型，我创建了kafka producer并调用了不同的方法，但不确定什么是正确的编程方法，这样流就不会中断，性能就不会受到影响。请帮忙。
模型1：

for(int i=1; i < 100; i++){
    Producer<String, String> producer = new KafkaProducer<String, String>(props);

    ProducerRecord<String, String> data = new ProducerRecord<String, String>(
        topicName, 
        String.valueOf(i)
    );

    producer.send(data);
    producer.close();
}

模型2：

Producer<String, String> producer = new KafkaProducer<String, String>(props);
for(int i=1; i < 100; i++) {

    ProducerRecord<String, String> data = new ProducerRecord<String, String>(
        topicName, 
        String.valueOf(i)
    );

    producer.send(data);
    producer.close();
}

模型3：

Producer<String, String> producer = new KafkaProducer<String, String>(props);
for(int i=1; i < 100; i++){

    ProducerRecord<String, String> data = new ProducerRecord<String, String>(
        topicName, 
        String.valueOf(i)
    );

    producer.send(data);
}
producer.close();

模型4：

for(int i=1; i < 100; i++){
    Producer<String, String> producer = new KafkaProducer<String, String>(props);

    ProducerRecord<String, String> data = new ProducerRecord<String, String>(
        topicName,
        String.valueOf(i)
    );

    producer.send(data);
    producer.flush();
    producer.close();
}

模型5：

Producer<String, String> producer = new KafkaProducer<String, String>(props);
for(int i=1; i < 100; i++){
    ProducerRecord<String, String> data = new ProducerRecord<String, String>(
        topicName, 
        String.valueOf(i)
    );

    producer.send(data);
    producer.flush();
    producer.close();
}

模型6：

Producer<String, String> producer = new KafkaProducer<String, String>(props);
for(int i=1; i < 100; i++){

    ProducerRecord<String, String> data = new ProducerRecord<String, String>(
        topicName, 
        String.valueOf(i)
    );

    producer.send(data);
    producer.flush();
}
producer.close();